Pricing and limits

Running compute (vCPU-seconds plus GiB-seconds), paused storage, runtime instances, storage and egress, and managed Postgres or Redis. The full list is on Pricing and limits.

miosa usage my-box           # runtime, time left and estimated cost

Storage only. Billing for compute stops when the sandbox leaves running. See Sizing and limits.

xs (1 vCPU, 2 GiB, 10 GiB) through xl (16 vCPU, 32 GiB, 80 GiB), with small as the default. See Sizing and limits.

Concurrent sandboxes per plan (Developer 10, Business 250, Enterprise 500; upgrade or raise it through support), timeout_sec 1 to 86400, an 86,400 second maximum exec timeout, a 500 MB maximum upload, and a 1 hour browser token TTL. See Pricing and limits.

60 to 30,000 requests per minute per API key by tier, and 1,000/25,000/100,000 per minute per workspace on Developer/Business/Enterprise, with 429 and Retry-After on overrun. See Rate limits.

The Quotas API sets max_sandboxes, max_concurrent, max_storage_gb, and max_credit_cents per external_user_id, enforced at create time.

MIOSA bills you once; you charge your customers using attribution. The Usage Rollup API groups by external_user_id, external_project_id, or workspace_id. See Usage and billing.

It depends on the grant. Purchased credits (top-ups and auto-recharge) do not expire. Recurring plan credits expire at the start of the UTC day after the 30-day billing period. Promotional credits (the one-time welcome grant and promo-code redemptions) expire 180 days after they are granted. The nearest expiry is reported as credit_expiry_at on GET /api/v1/credits/balance. See Pricing and limits.